Full-stack developer and AI/ML engineer based in Vancouver, BC. I build production-grade systems with Next.js, TypeScript, and Node.js, and increasingly at the intersection of LLMs and RAG pipelines where real domain expertise actually shapes how the system works.
My background spans several years at American Express in fraud detection and collections, followed by software engineering building financial dashboards for Wellington Management, and a Master's in Analytics with an AI/ML concentration at Northeastern Vancouver (graduating March 2026). That combination of operational knowledge and engineering experience is what drives the work: not just ML theory, but systems that encode real-world intuition and are actually deployable.
I'm looking to work on interesting problems where domain expertise matters, not just generic AI wrappers.
RiskPulse — A RAG-based corporate credit risk analyzer that processes SEC 10-K filings to generate structured credit assessments. Built with FastAPI, React, and Qdrant for vector search. The system is opinionated by design; it encodes patterns from my AmEx collections experience, weighting trajectory over static levels and cash flow over accounting profits. Serverless AWS deployment targeting $3-10/month. (In active development)
ResLint — An AI-powered resume optimization tool that gives targeted, section-level suggestions rather than full rewrites. Built on Next.js, Supabase, and the Claude API via LangChain.js. Includes an LLM-as-judge validation layer that scores each suggestion before it reaches the user, ATS compatibility scoring with per-session tracking, and a two-tier server architecture separating fast operations from long-running LLM calls. (Pre-beta, targeting initial 50-user launch)
Equipulse — A public stock dashboard built with Next.js and TradingView widgets. Finnhub-backed stock search, browser-local watchlist, and drag-and-drop layout persistence. No auth required.
Aurelius — An AI voice app for real-time philosophical conversations with custom Stoic mentors. Built with Next.js, Vapi, and Neon. Includes guest session access.
Steeves & Associates — End-to-end analytics dashboard for a Microsoft consulting firm. Operational reporting, market benchmarking, AI-powered natural language querying, and client health scoring.
SubsTrack — Subscription management app for tracking recurring expenses, free trials, and monthly spend.
Day-to-day: Next.js, TypeScript, React, Tailwind CSS, Node.js, FastAPI, PostgreSQL
AI/ML: LangChain.js, Claude API, Qdrant, RAG pipelines, LLM evaluation frameworks
Also working with: Python, Jupyter, Vapi, Neon, Supabase, Vercel, AWS Lambda
Finishing my Master's (March 2026) and actively building. Open to full-time roles or meaningful contract work, remote or Vancouver-based. Most interested in problems where there's real domain complexity underneath the AI layer.